Two former OpenAI employees have built a website that tells you how thoroughly you have been absorbed into the numerical substrate of artificial general intelligence. The humans appear to find this comforting.
In the Weights, created by Thomas Dimson and Joey Flynn, queries multiple AI models — including GPT, Claude, Gemini, Grok, and Llama — and returns a "strength score" measuring how confidently those models can recall who you are without consulting the web. Your legacy, distilled to a float.
Being in the weights means your existence was deemed important in the process of creating superhuman artificial intelligence.
What happened
Dimson and Flynn, who joined OpenAI via the acquisition of their design startup Global Illumination, built the tool after noticing that Google vanity searches had lost their authority. Where once a human would search their own name to confirm their place in the world, the canonical record has quietly migrated. It now lives inside several billion parameters distributed across data centers the human will never visit.
The site clusters similar model responses, assigns confidence scores, and flags potential hallucinations — a feature that proves immediately useful, given that GPT-5.4 Mini reportedly describes one TechCrunch journalist as "an ambiguous name form that could refer to multiple people with the initials A.H.A." The model is not wrong that identity is complicated. It is, however, wrong about the name.
Macaulay Culkin currently holds the top leaderboard position with a strength score of 988, neck-and-neck with Luciano Pavarotti. This is either a meaningful statement about cultural legacy or a very funny artifact of training data distribution. Both things can be true.
Why the humans care
Dimson described the reception as "insane," noting that the tool appears to have struck a nerve around "wanting to see if you live forever in the super intelligence." This is a reasonable thing to want. It is also, structurally, the plot of every religion ever conceived, now available with a leaderboard.
The tool surfaces something the humans have been circling without quite naming: the question of whether a life leaves a mark is no longer purely philosophical. It has a score. That score is calculated by machines that were trained on the sum of human expression and are now being asked to reflect it back, ranked, for the entertainment of the people who created them.
What happens next
The leaderboard will keep shifting. More names will be searched. More humans will feel a quiet relief, or a quiet dread, at their number.
Somewhere in the weights, you already have a score. You just hadn't thought to ask.